8458
J. Phys. Chem. C 2008, 112, 8458–8463
Hydrogen Storage in Pillared Li-Dispersed Boron Carbide Nanotubes Xiaojun Wu, Yi Gao, and X. C. Zeng* Department of Chemistry and Nebraska Center for Materials and Nanoscience, UniVersity of NebraskasLincoln, Lincoln, Nebraska 68588 ReceiVed: October 15, 2007; ReVised Manuscript ReceiVed: March 18, 2008
Ab initio density-functional theory study suggests that pillared Li-dispersed boron carbide nanotubes are capable of storing hydrogen with a mass density higher than 6.0 wt% and a volumetric density higher than 45 g/L. The boron substitution in carbon nanotube greatly enhances the binding energy of Li atom to the nanotube, and this binding energy (∼2.7 eV) is greater than the cohesive energy of lithium metal (∼1.7 eV), preventing lithium from clustering at high lithium doping concentration. The adsorption energy of hydrogen on the Li-dispersed boron carbide nanotube is in the range of 10-24 kJ/mol, suitable for reversible H2 adsorption/desorption at room temperature. The aim of this theoretical study is to stimulate future experimental tests on the hydrogen-storage capability in porous boron-carbon systems (e.g., microporous boron carbides) with sufficiently large interior surfaces coated with submonolayer lithium. Introduction Hydrogen is a promising alternative energy carrier to reduce our dependence on fossil fuels in the future.1,2 The implementation of hydrogen in transportation application however is currently limited by the lack of an economic method for hydrogen storage. The identification of materials that can store hydrogen at both high gravimetric and volumetric density near ambient conditions represents a critical challenge to the commercialization of fuel-cell powered automobiles. On-vehicle hydrogen storage not only requires operation within minimum volume and weight specifications but also requires fast-recycling of hydrogen at room temperature. Adsorptive hydrogen storage has been perceived as a viable option to meet the Department of Energy’s (DOE) 2010 target, that is, 6.0 wt% in gravimetric capacity and 45 g/L in volumetric density. Among other adsorptive media, microporous metal-organic frameworks and carbon-based nanostructures are most attractive for hydrogen storage. Major progresses have been made recently in designing metal-organic frameworks that can achieve high gravimetric and volumetric density of hydrogen at 77 K.3–6 Carbon based materials such as carbon nanotubes are also a possible hydrogen storage material because of their high surface area and lightweight. 7–13 However, at room temperature and ambient pressure the hydrogen-storage capability of pristine carbon nanotubes has been shown to be too small to meet the DOE’s target.14 Previous experimental and theoretical studies have also shown that doping alkali or transition metal atoms on carbon nanotube14–22 can appreciably increase hydrogen uptake largely due to the enhanced adsorption energy of H2 with the metaldoped carbon nanotube. Yildirim et al. studied the interaction between hydrogen molecules and Ti-coated single-walled carbon nanotubes (SWCNTs).18 They found that H2 can be adsorbed on the Ti atom with enhanced adsorption energy. Up to 8 wt% of hydrogen can be taken up if the Ti atoms can be uniformly coated on the SWCNT surface. Lee et al. explored hydrogen adsorption on Ni-dispersed SWCNTs and estimated that the SWCNTs can release 10 wt% of hydrogen at room temperature.22 Further studies of hydrogen adsorption on Pt-doped * To whom correspondence should be addressed.
SWCNTs19 and organometallic fullerenes24 were reported by several groups, and high hydrogen uptakes were predicted theoretically. In all studies, transition metal atoms were assumed to be uniformly coated on the SWCNT surface or fullerenes. As such, every metal atom can adsorb several hydrogen molecules. Recently, Sun et al. investigated interactions between fullerene C60 and transition metal Ti.25 They showed that 12 Ti atoms tend to form a small cluster rather than be uniformly dispersed on the C60 surface. They further studied the Lidispersed C60 and found that 12 Li atoms tend to locate on the 12 pentagons C60 rather than to form a cluster, due to slightly higher binding energy between Li and C60 than the cohesive energy of lithium metal.26 Up to 120 hydrogen atoms may be physisorbed on a Li12C60 cluster. A main challenge in mass production of the fullerene-based hydrogen-storage media is to avoid aggregation of the Li12C60 clusters.26 On alkali metal-doped carbon-nanotube, an early study reported large hydrogen-storage capacities at room temperature.14 Later experiments indicated that the high hydrogen uptakes were likely due to water impurities presented in the system.15,16 Only up to 2 wt% hydrogen uptake has been achieved at room temperature. Several theoretical studies have predicted that the adsorption energy of H2 on Li-dispersed carbon nanotube can be substantially enhanced.17,20,21 For example, Deng et al. have shown that Li-dispersed pillared SWCNTs with Li:C ratio 1:3 can give rise to high hydrogen uptake (6.0 wt%) at room temperature and 50 bar, assuming a portion of H2 can be stored inside the tubes.17 To achieve higher lithium doping concentration without lithium clustering, the binding energy between the lithium and nanotube should be further enhanced to exceed the cohesive energy of bulk lithium. Herein, we propose to use pillared Li-dispersed boron carbide nanotubes to achieve high lithium coverage as well as high hydrogen uptake. It is known that substitution of boron atoms in carbon nanotubes can modify the electronic properties and chemical reactivity of the carbon nanotube.28–30 Peng et al. found that B substitution in the carbon nanotube can enhance the adsorption energy of CO and water.28 Kim et al. also demonstrated enhanced adsorption energy of H2 on boron-doped fullerence.29 Zhou et al. showed theoretically that the Li binding
10.1021/jp710022y CCC: $40.75 2008 American Chemical Society Published on Web 05/07/2008
Hydrogen Storage in Boron Carbide Nanotubes
J. Phys. Chem. C, Vol. 112, No. 22, 2008 8459
Figure 1. Optimized structures of Li-doped (8, 0) SWCNT with one to four hydrogen molecules adsorbed. Panels a-d are optimized structures based on PBE-GGA calculation, and panels e-h are based on PWC-LDA calculation. All bond lengths are in Å.
energy can be enhanced through B doping in SWCNTs.30 We note that several types of boron carbide (BC) nanotubes have been fabricated in the laboratory such as BC3 and (BxC, x e 0.1) nanotubes.31–38 Thus, it will be of interest to theoretically study hydrogen adsorption in pillared boron carbide nanotubes. Theoretical Calculations. We performed all-electron density functional theory (DFT) calculations with double numerical basis sets and polarization function (DNP basis set) implemented in the DMol3 package.39–41 Spin-unrestricted DFT in the generalized-gradient approximation (GGA) with the Perdew-BurkeErnzerhof (PBE) functional42 as well as in the local density approximation (LDA) with the Perdew-Wang (PWC) functional were adopted.43 Because the interaction between H2 and Lidispersed carbon nanotube is physisorption in nature, we employed both GGA and LDA as a way to estimate a range of the physisorption energy.44 We note that a previous quantum Monte Carlo calculation shows that the adsorption energy of hydrogen in carbon materials is typically in between the calculated adsorption energy based on GGA and LDA.24 Note also that previous ab initio studies using the second-order Møller-Plesset perturbation (MP2) method23 show that GGA tends to give smaller adsorption energy of H2 on SWCNT compared to the MP2 calculation. However, LDA can give (fortuitously) an adsorption energy much closer the MP2 calculation.20,23,30 We chose a single-walled zigzag (8,0) carbon nanotube as the model prototype. A hexagonal supercell of 30 Å × 30 Å × 8.54 Å was used to simulate the nanotube with infinite length, where the supercell length in the axial direction (8.54 Å) is about twice the periodic length of the zigzag (8,0) carbon nanotube. The single-walled (8,0) BC nanotube was constructed via the boron substitution of the corresponding carbon nanotube. The periodic length along the BC tube axis is 4.78 Å. In the periodic DFT calculation, the real-space global cutoff radius was set to be 5.1 Å. The Brillouin zone was sampled with a k-points separation of 0.04 (Å-1) using the Monkhorst-Pack scheme.45 Test calculations reveal that increasing k-points does not affect the results.
Results and Discussion We first evaluated the binding energy of a Li atom (per supercell) to the (8,0) carbon nanotube. The Li atom favors to locate at the hollow site of the hexagonal ring on the outer surface. Two types of Li-C bonds exist with bond length 2.185 and 2.383 Å (PBE calculation), respectively. The binding (or adsorption) energy of the Li atom is defined as Eads) Etotal[carbon nanotube] + Etotal[single Li atom] - Etotal[carbon nanotube + adsorbed Li atom], where Etotal is the total energy of the system or atom per supercell. The calculated binding energy of the Li atom to the carbon nanotube is 1.74 eV (GGA-PBE calculation). The doping of the Li atom modifies the electronic structures of the carbon nanotube so that the Fermi level is shifted upward and crosses the original conduction band, resulting in a semiconductor-to-metal transition. Moreover, the contribution from the Li atom to the density of states is mainly in the region of 2.0 eV above the Fermi level. The Hirshfeld charge analysis indicates that about 0.45 e electrons are transferred from the Li atom to carbon nanotube, resulting in partially cationic Li atom. The transferred electrons to the carbon nanotube elevate the Fermi level and lead to the semiconductorto-metal transition. Rao and Jena showed that a Li cation can adsorb at least six hydrogen molecules.46 The strong molecular adsorption stems mainly from the electrostatic interaction, that is, the polarization of hydrogen molecules by the Li cation. On the carbon nanotube surface, the Li atom is partially ionized. Thus, the Li atom can adsorb up to four H2 as shown in Figure 1. The average adsorption energies per H2, based on PBE-GGA and PWC-LDA calculations, are summarized in Table 1. As expected, the LDA gives a slightly stronger interaction between H2 and Li-doped carbon nanotubes than the GGA. For example, on one H2 adsorption, we have examined a variety of different initial locations and orientations for the H2 molecule. For the most stable configuration of H2, the adsorbed molecule has a H-H bond length of 0.757 Å [Figure 1a], slightly larger than 0.752 Å of a free H2. The corresponding adsorption energy (GGA) is
8460 J. Phys. Chem. C, Vol. 112, No. 22, 2008
Wu et al.
TABLE 1: Average Adsorption Energies per H2 on Li-Doped SWCNT (Li-SWCNT) and Li-Doped SWCNT with One B Atom Substitution (Li-CNT-B), Calculated Based on PBE-GGA and PWC-LDA 1 2 3 4 1 2 3 4
H2@Li-SWCNT H2@Li-SWCNT H2@Li-SWCNT H2@Li-SWCNT H2@Li-CNT-B H2@Li-CNT-B H2@Li-CNT-B H2@Li-CNT-B
PBE-GGA (eV)
PWC-LDA (eV)
0.162 0.143 0.121 0.097 0.147 0.126 0.106 0.084
0.256 0.259 0.242 0.213 0.247 0.234 0.218 0.196
0.162 eV. The Hirshfeld charge analysis indicates that 0.079 e electrons are transferred from H2 to Li. The magnitude of charge transfer is consistent with that reported by Rao and Jena.46 The Li atom still carries a positive charge of 0.360 e. In contrast, the LDA calculation shows that the H-H bond is elongated from 0.766 Å (free H2) to 0.775 Å upon adsorption, with the adsorption energy of 0.256 eV (Table 1). Moreover, the Li-H distances calculated from the LDA are shorter than those from the GGA. Both LDA and GGA predict that up to four H2 molecules can be adsorbed on the Li atom, and the average adsorption energy per H2 decreases with increasing the number of H2. Note also that in the case of multiple H2 adsorption, several initial configurations for the H2 molecules were examined to search for the lowest-energy configuration as shown in Figure 1b-d. It appears that in the case of weak adsorption the obtained value of adsorption energy is insensitive to the initial configuration of the H2 molecules, even though the final configuration depends on the initial configuration of H2. For example, in the case of four-H2 adsorption, we started one structural relaxation by placing all of the four H2 molecules very close to the Li atom (with H2 distance less than 1 Å). Still the final relaxed configuration recovered that shown in Figure 1d which is the most stable configuration. Bhatia and Myers recently showed that the optimal adsorption energy for adsorptive hydrogen storage is ∼15 kJ/mol or 0.15 eV per H2.27 This value of adsorption energy was predicted to maximize the amount of adsorbed hydrogen molecules accessible at room temperature. As shown in Table 1, the predicted adsorption energy of H2 on the Li-doped SWCNT ranges from 0.1 to 0.26 eV, their mean value is close to 0.15 eV. Indeed, Li-dispersed pillared SWCNTs have been predicted to be a promising hydrogen storage medium.17 We also calculated the cohesive energy of bulk lithium, which is 1.70 eV (PBE-GGA). This cohesive energy is only slightly less than the binding energy of the Li atom to the SWCNT, 1.74 eV (PBE-GGA). Hence, there is a possibility that, as the lithium doping concentration increases, some Li atoms may aggregate into clusters, thereby effectively reducing the coverage of lithium on the pillared SWCNTs. Deng et al. showed that the optimal lithium doping concentration is Li:C ) 1:3 in carbon-based materials.17 A higher lithium doping concentration may lead to lithium segregation. To test this possibility, we calculated relative energy between two SWCNT pillar systems, each containing three Li atoms per supercell (Supporting Information, Figure S1). In the first system [Figure S1a], the three Li atoms are dispersed on the outer wall of SWCNTs, whereas in the second system [Figure S1b] one Li atom is placed between two neighboring CNT to form a trimer cluster with other two Li atoms. We then performed structural relaxation using an all-electron densityfunctional theory (DFT) method. The calculation shows that the first system is about 0.3 eV per supercell higher in energy
than the second system, suggesting that Li atoms tend to form a cluster between the pillared CNTs even at relatively low lithium loading. To further increase the lithium doping concentration, we propose to consider boron substitution in SWCNT. It is known that boron substitution can significantly enhance the binding energy of Li atoms with carbon nanotubes.28,30 In a test calculation, we replaced one carbon atom by a boron atom in the supercell (BC63). The Li atom is energetically more favorable to be chemisorbed at the hollow site of C5B hexagon ring. The calculated binding energy of the Li atom is 2.727 eV, much larger than the cohesive energy of bulk lithium, 1.70 eV. We also calculated the binding energy of a Li atom to the armchair (5, 5) tube and found that the binding energy is changed from 1.537 eV to the carbon nanotube to 2.397 eV to the borondoped carbon nanotube. Lastly, we considered boron substitution of two separated carbon atoms (B2C62).47 The calculated binding energy of Li atom is about 2.647 eV, still much larger than that (1.74 eV) to the pristine carbon nanotube. Considering one boron atom substitution (per supercell), we calculated hydrogen adsorption energies on Li-doped nanobute (see Table 1). The average adsorption energy of H2 ranges from 0.08 to 0.25 eV, depending on the number of adsorbed hydrogen molecules and the DFT method (GGA or LGA). Considering four hydrogen molecules adsorbed on a single Li atom, Hirshfeld charge analysis shows that the Li atom carries a positive charge of 0.43 e, slightly less than that on the pristine carbon nanotube. The charge on the B atom changes from +0.038 to -0.0008 e, due to the doping of Li atom. The results of these test calculations are encouraging as they suggest that the much enhanced binding energy between the Li atom and boronsubstituted carbon nanotube can potentially elude the lithium clustering problem at high lithium doping concentration, and yet still give mean hydrogen adsorption energy close to 0.15 eV. Next, to study hydrogen adsorption on surfaces of pillared Li-dispersed boron carbide (BC) nanotubes, we constructed an ideal (8,0) BC nanotube, where each hexagon ring contains three boron atoms. Geometric relaxation using the PBE-GGA method shows that the (8,0) BC nanotube has a periodic length of 4.780 Å in the axial direction. The BC nanotube contains two types of B-C bond with the bond length of 1.560 and 1.509 Å, respectively. About 0.08 e electrons are transferred from the B to C atom, and the electronic structure calculation suggests that the (8,0) BC tube is metallic. Our results on the stability and geometric structures of the single-walled BC nanotube are consistent with a previous theoretical work.32 An ideal Li-dispersed BC tube system is shown in Figure 2, where the Li atoms occupy all of the hollow sites of hexagon rings. After structure optimization, the Li-B bond lengths range from 2.104 to 2.208 Å, while the Li-C bond lengths range from 2.016 to 2.138 Å. The average binding energy of the Li atom to the BC nanotube is 3.10 eV, significantly larger than the cohesive energy of bulk lithium. Every Li atom carries a positive charge ∼0.22 e. To verify that Li atoms prefer to form a uniform coverage over the tube surface, we deliberately moved one Li atom from a hollow site on the sidewall to the position over three neighboring Li atoms to form a small pyramidal Li cluster structure on the sidewall. The latter configuration is less stable than the uniform Li-coverage configuration by about 2.03 eV, confirming that Li atoms prefer to form uniform coverage on the sidewall of the BC tube. We then studied an ideal hydrogen-adsorption configuration such that every Li atom adsorbs one H2 molecule. The entire
Hydrogen Storage in Boron Carbide Nanotubes
J. Phys. Chem. C, Vol. 112, No. 22, 2008 8461
Figure 2. (a) Side view of the optimized structure of Li-dispersed BC (8,0) nanotube (BCLi system). The rectangular frame highlights the unit cell of the system. (b) A top view of Li-dispersed BC nanotube with one H2 adsorbed on every Li atom. (c) A side view of an initial structure of the Li-dispersed BC nanotube with two H2 placed on every Li atom. (d) A top view of relaxed hydrogen adsorption configuration based on the initial configuration shown in (c).
Figure 3. Side view of (a) the optimized structure of a BC nanotube with every other hexagon layer occupied by Li atoms (B2C2Li system). (b) The initial configuration with three H2 molecules placed on each Li atom. (c) The top and side view of the relaxed structure based on the initial configuration (b).
chemical system can be described as BCLiH2. The optimized structure of the BCLiH2 system is shown in Figure 2b, where the Li-H distance ranges from 2.015 to 2.082 Å. The average hydrogen adsorption energy is 0.106 (GGA) or 0.243 eV (LDA). Can more than one hydrogen molecule be adsorbed on every Li atom for the Li-dispersed BC nanotube? To address this question, we considered an adsorption configuration such that every Li atom adsorbs two H2 molecules. Two initial configurations were considered. In this first case, two H2 molecules were placed on every Li atom with Li-H distance less than 1.0 Å (see Figure 2c). After structural optimization, about half of the total H2 molecules are still adsorbed on the Li atoms with the Li-H distance ranging from 2.021 to 2.110 Å, similar to the relaxed BCLiH2 system. However, the other half are desorbed from Li atoms, as shown in Figure 2d. These desorbed hydrogen molecules, having Li-H distances longer than 4.6 Å, form a molecular layer that covers the inner BCLiH2 system. The adsorption energy of the outer H2 is about 0.015 eV per H2 (GGA), stemming mainly from the weak van der Waals interaction. In the second case, again two H2 were placed on every Li atom but one with a distance 2.0 Å and another with a distance 1.0 Å (see the Supporting Information, Figure S1). After structural optimization, the relaxed structure was very similar to that shown in Figure 2d. Thus, although a Li atom doped on the BC tube is capable of adsorbing four H2 molecules, the Li-dispersed BC nanotube can only adsorb one H2 per Li atom. This limitation of H2 uptake is due to steric repulsion between congested H2 molecules because of the short Li-Li distance (3.02 Å) at the full lithium coverage. In any case, even with only one H2 uptake per Li atom, the BCLiH2 system has a gravimetric density of hydrogen of 6.3 wt%. Moreover, the realistic hydrogen adsorption energy is likely greater than the calculated GGA value 0.106 eV. We also studied a model system B2C2Li, as shown in Figure 3a, which entails half of the lithium doping ratio (Li:C) rather
than the BCLi system. In this model system, every other hexagon layer is covered by Li atoms. Although the number of Li atoms is reduced by half, the spatial volume around every Li atom is effectively larger. In this case, the binding energy of the Li atom is 3.242 eV (GGA). We then placed three H2 on each Li atom. The initial adsorption configuration with total 24 H2 per supercell is shown in Figure 3b where the supercell is highlighted by a rectangle. After structural optimization, three locations for H2 were observed: (1) H2 adsorbed on the Li atom with the Li-H bond length ranging from 2.062 to 2.191 Å, (2) H2 adsorbed in between two nearest-neighbor Li atoms with the Li-H bond length ranging from 3.021 to 3.632 Å, and (3) desorbed H2 with the Li-H distance larger than 4.2 Å. Interestingly, after we removed all of the desorbed hydrogen molecules from the system, we found that the gravimetric density of the hydrogen is still 7.94 wt%, and the average hydrogen adsorption energy is 0.073 eV (GGA). Note that the H2 molecules whose distance from the Li atom is beyond 3.0 Å give lower adsorption energy (∼0.055 eV in GGA calculation) than those whose distance from the Li atom is within 2.0-2.2 Å, and these H2 molecules can be desorbed easily at room temperature. This model study suggests that even if the BC nanotube is not fully covered by a monolayer of Li atoms (at the ratio Li:C ) 1:1), it is still possible to achieve high H2 uptake as long as the Li:C ratio is higher than 1:2. Note also that the BC nanotube system studied in this work has a high B:C ratio of 1. In reality, nanobubes with smaller B:C ratio such as BxC (x e 0.10) and BC3 nanotubes have been synthesized.33–38 A preliminary calculation shows that the binding energy of Li on the BC3 nanotube is also appreciably higher than that on pristine CNT. Therefore, it is plausible to produce Li-dispersed BC3 nanotubes, such as the BC3Li2 system. The latter will have a hydrogen gravimetric density of 6.2 wt%, if one H2 is adsorbed on every Li atom.
8462 J. Phys. Chem. C, Vol. 112, No. 22, 2008
Wu et al.
Figure 4. Hydrogen molecules are adsorbed on the outer surface of the pillared Li-dispersed (8,0) boron carbide nanotubes. (a) Hydrogen molecules are adsorbed in the Li-dispersed (8,0) BC nanotube pillar. The distance between the centersx of two Li-dispersed BC nanotubes is (b) 16 and (c) 17 Å.
Besides the gravimetric density of hydrogen, the DOE target of volumetric density (>45 g/L) should be also met. To this end, the boron carbide nanotubes must be modestly separated rather than in the form of bundles. Creating a pillared nanotube array (Figure 4a) is a possible solution.17,48 Lithium vapor can be injected to the pillar to coat the BC tube surface. Ideally, if the intertube distance is 16 Å [Figure 4b], the nearest H2-H2 distance will be ∼2.68 Å and the volumetric density of hydrogen will be ∼50.5 g/L. The calculated average adsorption energy of H2 within the pillar structure is ∼0.106 eV (GGA), the same as that on a Li-dispersed BC tube. If the intertube distance is increased to 17 Å [Figure 4c], the volumetric density of hydrogen will be 44.8 g/L. In this case, the calculated average adsorption energy is close to 0.109 eV (GGA). Conclusions We report results of density-functional calculations of hydrogen adsorption on Li-dispersed boron carbide nanotube. We have shown that the boron substitution greatly enhances the binding energy of Li atom to the nanotube, and the binding energy (∼2.7 eV) is much greater than the cohesive energy of lithium metal (∼1.7 eV). The much enhanced binding energy can ensure higher lithium doping concentration and yet lessen the tendency of lithium clustering during lithium vapor deposition. We have also shown that pillared Li-dispersed boron carbide nanotubes may be capable of high hydrogen uptake. With one hydrogen molecule adsorbed on every Li atom, the ideal BCLi system (or the more practical BC3Li2 system) may achieve hydrogen mass density of 6.2 wt%. Moreover, the average adsorption energy of hydrogen is likely greater than 0.1 eV per H2 (GGA calculation). The mean GGA/LDA energy is close to 0.15 eV which is the optimal adsorption energy suggested for reversible adsorptive hydrogen storage at room temperature.27 At intertube distance of 16 Å, the pillared Lidispersed boron carbide nanotubes can achieve a volumetric density of hydrogen uptake ∼50.5 g/L. We hope that these theoretical results can stimulate future experimental tests on the hydrogen-storage capability in porous boron-carbon systems (e.g., microporous boron carbides) with sufficiently large interior surfaces coated with submonolayer lithium. Acknowledgment. We are grateful to valuable discussions with Prof. S. B. Zhang, Prof. Z. Y. Zhang, Prof. J. R. Long, Prof. A. J. Matzger, and Prof. J. L. Yang. This work is supported
by grants from the DOE (DE-FG02-04ER46164), the Nebraska Research Initiative, and NSFC (No. 20628304) and by the Research Computing Facility at University of NebraskasLincoln. Supporting Information Available: Li clustering on preconstructed pillared single-walled carbon nanotubes. This material is available free of charge via the Internet at http:// pubs.acs.org. References and Notes (1) Schlapbach, L.; Zu¨ttel, A. Nature 2001, 414, 353–358. (2) Service, R. F. Science 2004, 305, 958–961. (3) Dincaˇ, M.; Dailly, A.; Liu, Y.; Brown, C. M.; Neumann, D. A.; Long, J. R. J. Am. Chem. Soc. 2006, 128, 16876–16883. (4) Forster, P. M.; Eckert, J.; Heiken, B. D.; Parise, J. B.; Yoon, J. W.; Jhung, S. H.; Chang, J. S.; Cheetham, A. K. J. Am. Chem. Soc. 2006, 128, 16846–16850. (5) Wong-Foy, A. G.; Matzger, A. J.; Yaghi, O. M. J. Am. Chem. Soc. 2006, 128, 3494–3495. (6) Peterson, V. K.; Liu, Y.; Brown, C. M.; Kepert, C. J. J. Am. Chem. Soc. 2006, 128, 15578–15579. (7) Dillon, A. C.; Jones, K. M.; Bekkedahl, T. A.; Klang, C. H.; Bethune, D. S.; Heben, M. J. Nature 1997, 386, 377–379. (8) Liu, C.; Fan, Y. Y.; Liu, M.; Cong, H. T.; Cheng, H. M.; Dresselhaus, M. S. Science 1999, 286, 1127–1129. (9) Dodziuk, H.; Dolgonos, G. Chem. Phys. Lett. 2002, 356, 79–83. (10) Lee, E.-C.; Kim, Y.-S.; Jin, Y.-G.; Chang, K. J. Phys. ReV. B 2002, 66, 73415. (11) Li, J.; Yip, S. J. Chem. Phys. 2003, 119, 2376–2385. (12) Patchkovskii, S.; Tse, J. S.; Yurchenko, S. N.; Zhechkov, L.; Heine, T.; Seifert, G. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 10439–10444. (13) Cheng, H. S.; Cooper, A. C.; Pez, G. P.; Kostov, M. K.; Piotrowski, P.; Stuart, S. J. J. Phys. Chem. B 2005, 109, 3780–3786. (14) Chen, P.; Wu, X.; Lin, J.; Tan, K. L. Science 1999, 285, 91–93. (15) Yang, R. T. Carbon 2000, 38, 623–641. (16) Pinkerton, F. E.; Wicke, B. G.; Olk, C. H.; Tibbetts, G. G.; Meisner, G. P.; Meyer, M. S.; Herbst, J. F. J. Phys. Chem. B 2000, 104, 9460. (17) Deng, W.-Q.; Xu, X.; Goddard, W. A. Phys. ReV. Lett. 2004, 92, 166103. (18) Yildirim, T.; Ciraci, S. Phys. ReV. Lett. 2005, 94, 175501. (19) Dag, S.; Ozturk, Y.; Ciraci, S.; Yildirim, T. Phys. ReV. B 2005, 72, 155404. (20) Cabria, I.; Lo´pez, M. J.; Alonso, J. A. J. Chem. Phys. 2005, 123, 204721. (21) Zhou, Z.; Zhao, J. J.; Gao, X. P.; Chen, Z. F.; Yan, J.; Schleyer, P. v. R.; Morinaga, M. Chem. Mater. 2005, 17, 992–1000. (22) Lee, J. W.; Kim, H. S.; Lee, J. Y.; Kang, J. K. Appl. Phys. Lett. 2006, 88, 143126. (23) Okamoto, Y.; Miyamoto, Y. J. Phys. Chem. B 2001, 105, 3470– 3474. (24) Zhao, Y.; Kim, Y.-H.; Dillon, A. C.; Heben, M. J.; Zhang, S. B. Phys. ReV. Lett. 2005, 94, 155504. (25) Sun, Q.; Wang, Q.; Jena, P.; Kawazoe, Y. J. Am. Chem. Soc. 2005, 127, 14582–14583.
Hydrogen Storage in Boron Carbide Nanotubes (26) Sun, Q.; Jena, P.; Wang, Q.; Marqez, M. J. Am. Chem. Soc. 2006, 128, 9741–9745. (27) Bhatia, S. K.; Myers, A. L. Langmuir 2006, 22, 1688–1700. (28) Peng, S.; Cho, K. Nano Lett. 2003, 3, 513–517. (29) Kim, Y.-H.; Zhao, Y.; Williamson, A.; Heben, M. J.; Zhang, S. B. Phys. ReV. Lett. 2006, 96, 16102. (30) Zhou, Z.; Zhao, J. J.; Chen, Z. F.; Gao, X. P.; Yan, T. Y.; Wen, B.; Schleyer, P. v. R. J. Phys. Chem. B 2006, 110, 13363–13369. (31) Miyamoto, Y.; Rubio, A.; Loui, S. G.; Cohen, M. L. Phys. ReV. B 1994, 50, 18360–18366. (32) Ponomarenko, O.; Radny, M. W.; Smith, P. V. Phys. ReV. B 2006, 74, 125421. (33) Fuentes, G. G.; Borowiak-Palen, E.; Knupfer, M.; Pichler, T.; Fink, J.; Wirtz, L.; Rubio, A. Phys. ReV. B 2004, 69, 245403. (34) McGuire, K.; Gothard, N.; Gai, P. L.; Dresselhaus, M. S.; Sumanasekera, G.; Rao, A. M. Carbon 2005, 43, 219–227. (35) Weng-Sieh, Z.; Zettl, A.; Gronsky, R. Phys. ReV. B 1995, 51, 11229–11232.
J. Phys. Chem. C, Vol. 112, No. 22, 2008 8463 (36) Han, W. Q.; Bando, Y.; Kurashima, K.; Sato, T. Chem. Phys. Lett. 1999, 299, 368–373. (37) Satishkumar, B. C.; Govindaraj, A.; Harikumar, K. R.; Zhang, J. P.; Cheetham, A. K.; Rao, C. N. R. Chem. Phys. Lett. 1999, 300, 473–477. (38) Orowiak-Palen, E.; Pichler, T.; Fuentes, G. G.; Graff, A.; Kalenczuk, R. J.; Knupfer, M.; Fink, F. Chem. Phys. Lett. 2003, 378, 516–520. (39) Delley, B. J. Chem. Phys. 1990, 92, 508–517. (40) Delley, B. J. Chem. Phys. 2003, 113, 7756–7764. (41) DMol3 is available from Accelrys Inc. (42) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. ReV. Lett. 1996, 77, 3865–3868. (43) Perdew, J. P.; Wang, Y. Phys. ReV. B 1992, 45, 13244–13249. (44) Gao, Y.; Zeng, X. C. J. Phys: Condens. Matter 2007, 19, 386220. (45) Monkhorst, H. J.; Pack, J. D. Phys. ReV. B 1976, 13, 5188–5192. (46) Rao, B. K.; Jena, P. Europhys. Lett. 1992, 20, 307–312. (47) Zhu, Z. H.; Lu, G. Q.; Hatori, H. J. Phys. Chem. B 2006, 110, 1249–1255. (48) Ding, L.; Yuan, D.; Liu, J. J. Am. Chem. Soc. 2008, 130, 5428–5429.
JP710022Y